5-Gram Unified Event Model
In this paper, a 5-gram unified event model (UEM5) is presented, which facilitates user session identification and provides better quality data for mining algorithms. Experimental results show that UEM5 is a high quality data source and has favorable online logging and offline mining performance. It can be well applied in E-commerce applications, including Web personalization, recommendation and business intelligence.
Shengyi Li, Yingchun Liu, Rongbo Zhu, Hongguang Li, Wensi Ding
X. Y. Zou "5-Gram Unified Event Model", Applied Mechanics and Materials, Vols. 34-35, pp. 1319-1322, 2010